On 20 Jul 2010 13:03:57 +0100
"N.M. Maclaren" wrote:
>
> Not on most systems. While this is more clearly illegal, similar
> remarks apply to its safety. If there were any debugging C compilers
> around, it might well get trapped, but those are about as common as
> unicorns.
>
> It's a horrible
Jeff -
I think falling back to GCC built-in if available is a rational idea. We've
been using them in another project without any problems. They are potentially
a bit slower than the hand-crafted assembly because they generally use full
memory barriers when we only need read memory barriers,
One thing to note is that the GCC atomic intrinsics are not always
implemented either. Use of an intrinsic that is unimplemented in a
given GCC version for a given platform will result in an link failure
(trying to call an "external" implementation that probably does not
exist). So, even if t
*** This mail mainly targeted at Brian and George ***
Debian maintainer Manuel Prinz raised an idea to me this morning:
The Debian community compiles and tests Debian on a huge range of hardware
platforms. It's been a long-standing issue that Open MPI doesn't support all
of them (e.g., MIPS, A
On Jul 20 2010, Jeff Squyres wrote:
> Also, it seems like the 3rd parameter could be problematic if it ever
> goes larger than 2B -- it'll increment in the wrong direction, won't
> it?
Not on most systems.
Ah -- I just checked -- the associativity of + and (cast) are equal, and
are righ
On Jul 20, 2010, at 8:03 AM, N.M. Maclaren wrote:
> > Also, it seems like the 3rd parameter could be problematic if it ever
> > goes larger than 2B -- it'll increment in the wrong direction, won't it?
>
> Not on most systems.
Ah -- I just checked -- the associativity of + and (cast) are equal,
On Jul 20 2010, Jeff Squyres wrote:
The change was to add casting:
} while (!OPAL_ATOMIC_CMPSET_32((int32_t*)&ep->eager_rdma_remote.seq,
(int32_t)ftr->seq, (int32_t)ftr->seq+1));
Is it safe to simply cast a (uint32_t*) to (int32_t*) in the first param?
Pretty safe. While there ARE
Please file CMR if this is needed in 1.4 and 1.5.
Thanks!
On Jul 20, 2010, at 2:31 AM, wrote:
> Author: cyeoh
> Date: 2010-07-20 02:31:16 EDT (Tue, 20 Jul 2010)
> New Revision: 23440
> URL: https://svn.open-mpi.org/trac/ompi/changeset/23440
>
> Log:
> Adds missing sys/stat.h include needed fo
The change was to add casting:
} while (!OPAL_ATOMIC_CMPSET_32((int32_t*)&ep->eager_rdma_remote.seq,
(int32_t)ftr->seq,
(int32_t)ftr->seq+1));
Is it safe to simply cast a (uint32_t*) to (int32_t*) in the first param?
Also, it seems like the 3rd
Thanks Chris!
On Jul 20, 2010, at 1:23 AM, Christopher Yeoh wrote:
> Hi Jeff,
>
> I've committed a fix for this in r23441
>
> Regards,
>
> Chris
>
> On Mon, 19 Jul 2010 11:58:31 -0400
> Jeff Squyres wrote:
>
>> Chris Yeoh --
>>
>> SVN blame says that this is your line of code. Can you fix
Hi Jeff,
I've committed a fix for this in r23441
Regards,
Chris
On Mon, 19 Jul 2010 11:58:31 -0400
Jeff Squyres wrote:
> Chris Yeoh --
>
> SVN blame says that this is your line of code. Can you fix?
>
>
> On Jul 17, 2010, at 12:27 PM, Ralph Castain wrote:
>
> > Yo IB-folks
> >
> > Are w
11 matches
Mail list logo
